Ending sentences with prepositons, is, of course, grammatically wrong.
Should be : Unfortunately, that is the state in which our society is.
Churchill showed just how silly this can get with the marvellous phrase "This is something up with which I shall not put."
Hehe. I showed my wife (English minor, German major) my post as I sent it saying, "I wonder what people might say". I guess I should have taken a leaf out of the pages of The Inquirer's play book and turned my [irony alert] into, "Warning: Blatant grammatical error and attempt at humour!!".
Hehe. I showed my wife (English minor, German major) my post as I sent it saying, "I wonder what people might say". I guess I should have taken a leaf out of the pages of The Inquirer's play book and turned my [irony alert] into, "Warning: Blatant grammatical error and attempt at humour!!".
Subtlety, much like proper grammar and punctuation, is an art lost in online forums.
I may have just missed his posts but I haven't seen Moki on these boards for a while. No comments on the MB rumors for dates and benchmarks. You don't suppose he has a test machine and is under a NDA do you?
It's fun to start rumors.
last time i mentioned that, he was very upset that i actually thought he had a ppc970 powermac to toy with.
(repeat after me) Essence of language is comprehension! Essence of language is comprehension! Essence of language is comprehension! Essence of language is comprehension!
(repeat after me) Essence of language is comprehension! Essence of language is comprehension! Essence of language is comprehension! Essence of language is comprehension!
May I revise that to "Essence of language is unambiguous comprehension"? I've see far too many cases where both sides of a discussion come away "understanding" the outcome, but both "understanding" very different things.
I've decided to post what I know about IBM's chip plans for the PPC and take some educated guesses as to how they relate to Apple.
Firstly, let me tell you why I know what I know. I have three good friends in the upper echelon of IBM. Two are in tech development positions of some kind, the other is in management. One is in Colorado, another in New York, and the last in North Carolina. I've gotten agreeing chip information from two of them. The third works in tech that doesn't involve chip manufacture. #3 has heard rumors, but hasn't been to official meeting or seen official documents.
Well, from what I understand, Apple came to IBM about two years ago requesting help because of Motorola's lack of interest in really building kick-butt processors for Apple. IBM agreed to help, although certain conditions were needed for this to happen. Most of this involved business related stuff and tech development costs, as well as some trade offs on technology developed by each company.
So, everyone knows about the 970 now, but there is also development on a new 750 based chip. There is already a 750 ready to go that is Altivec (I know, Moto terminology, just easier to know what it is that way) enhanced, can run with a FSB @ 400 or 450 MHZ with speeds up to 2GHZ. This is a modified 750fx, for all intents and purposes. These have been available for several months now, but Apple has not wanted them until now. It had to do with Apple's contractual obligations to Moto and the fact that they toast the G4 chip.
Apple is scheduled to have several palettes of these 750s (yes, up to 2 GHZ, lowest being 1.4) delivered to them in May. They also will have the 970s delivered to them @ speeds of 1.5 to 2.5. So, here's my educated guess:
The 970s now actually run more efficiently than when specs were announced, so that a high-end laptop could take a chip running at 1.8. I'm not sure if Apple will keep them all at 1.5 for power efficiency, but they may.
So, I think Apple introduces a completely revamped lineup in July. The iBooks and eMacs all use the 750 chip - iBooks running at 1.4, eMacs up to 2. The iMac may have a low-end model with this chip, and may be launched without the 970, giving the "Power" lineup a 3 to 6 month head start with the 970. I also don't think they'll call the 750 the G3. My bet is they may call it a Super G4 or a G4 Xtreme - something like that.
I see the PowerMacs getting an assortment of 970s, up to 2.5 with dual processor combos possible.
I think the 12" PB will get a 1.5 processor while the 15" PBs can get a 1.5 or 1.8 and the 17" comes with a 1.8
I think Apple may put 1MB of L3 on the 750 machines, 2MB on the PBs and PowerMacs. However, the PMs can take up to 4 - and it may be best for the laptops to only have 1.
I also know IBM is working on a 32bit new 750 based processor for Apple, so my guess is this will be used to delineate the iBook and eMac from the Power lineup. It may extend to the iMac too, but I'd guess a high-end iMac would have the 970.
And, now to the real kicker. Remember how I said that Apple had to make concessions to IBM? Well, one of them would allow IBM to manufacture and sell machines that ran Mac OS X - and would allow them sell it for servers, if a customer wanted it. This would not allow them to build laptops that ran the Mac OS, only desktop machines or servers.
Personally, I'm very skeptical, but hope some of it is true.
That is some thoroughly cool stuff. An altivec G3 smoking the G4, I'm SO willing to believe that (no joke here). I've always considered the G4 to be a dog. Also, this might explain the direction that will be taken with the i-line versus the pro-line. Making the i-line a line that is really appealing in terms of everything.
Nothing really new here... the poster could simply have read this forum and put together a "rumour" that fullfills what people are speculating about.
One mildly interesting idea in there, however, is that Apple wants to take legal action against Motorola but to do that they have to completely eliminate their dependency on Motorola products... and that they might hold back on the latest IBM chips until they can fullfill their contractual obligations and convert over completely. I'm skeptical, however. The timeframe on G3 w/ SIMD is aggressive and converting all entire lineups all at once seems too risky, especially when they just updated the eMacs. I don't buy it.
Moki has not posted since 4/21. His last post was in the 970 tower Mockup thread. Maybe he has just lost interest, but I really would like to hear what he has to say about all of these rampant rumors.
Nothing really new here... the poster could simply have read this forum and put together a "rumour" that fullfills what people are speculating about.
One mildly interesting idea in there, however, is that Apple wants to take legal action against Motorola but to do that they have to completely eliminate their dependency on Motorola products... and that they might hold back on the latest IBM chips until they can fullfill their contractual obligations and convert over completely. I'm skeptical, however. The timeframe on G3 w/ SIMD is aggressive and converting all entire lineups all at once seems too risky, especially when they just updated the eMacs. I don't buy it.
Yes, the "Apple wants to take legal action" rumors seem a stretch to me. Let me rephrase that, maybe Apple would like to take legal action, but the rumors saying they intend to seem a stretch.
The part of this post that appears the most far fetched comes in the last paragraph(re: as other people posted in Mac Forums). It's possible that Apple would agree to allow IBM to manufacture desktops w/ Mac OS X, but I don't think it very likely. However, maybe Apple learned something in their foray into the clones and have included some more restrictions, nawww, I'm dreaming, ain't happening. Oh well, we'll soon find out.
Still hoping for the IBM 850, that was the hook this post got to me with.(preposition that )
Macbidouille has stated that the new PM's will include Hypertransport. I haven't seen that addressed in this thread (although I'll admit that I haven"t read every post). Would someone much more knowledgeable than I comment on the ramifications of Hypertransport and the impact on throughput?
... in the end, good grammar is far more about good judgment, than it is about applying pre-set prescriptions: something that is completely lost in the way it's taught.
Macbidouille has stated that the new PM's will include Hypertransport. I haven't seen that addressed in this thread (although I'll admit that I haven"t read every post). Would someone much more knowledgeable than I comment on the ramifications of Hypertransport and the impact on throughput?
It depends on how it is used. The most likely use is to allow Apple to divide their chipset into two chips, one for I/O and one for the memory controller + FSB (+AGP?). That will help Apple rev the features of one without updating the other, and may allow them to share parts between high & low end lines.
Other uses of HT are harder to envision -- if they get a special GPU from ATI or nVidia (both of whom are part of the HT consortium) then it could replace AGP. This is much less likely, IMHO. The 970's FSB is HT-like but its not HT so this won't be used to communicate to the processor(s). Apple could also divide their chips further, but this tends to make the system more expensive so its not a benefit.
Really the use of HT is a technical detail that isn't particularly interesting to the end user.
It depends on how it is used. The most likely use is to allow Apple to divide their chipset into two chips, one for I/O and one for the memory controller + FSB (+AGP?). That will help Apple rev the features of one without updating the other, and may allow them to share parts between high & low end lines.
Other uses of HT are harder to envision -- if they get a special GPU from ATI or nVidia (both of whom are part of the HT consortium) then it could replace AGP. This is much less likely, IMHO. The 970's FSB is HT-like but its not HT so this won't be used to communicate to the processor(s). Apple could also divide their chips further, but this tends to make the system more expensive so its not a benefit.
Really the use of HT is a technical detail that isn't particularly interesting to the end user.
The initial standard of HT was a daisy chain connection and the comunication between slaves in a daisy chain was not allowed. All comunication between slaves has to be done through master.
Later on the consortium has added the possibility of a switch like RAPIDIO.
If 970 has HT the communication between processors could be done with a switch.
mercury computer systems says their new impact rt systems will be available this quarter, using a "next generation powerpc" processor with altivec. this should prove that the ppc970 will be available for some companies in may or june - and i'd bet apple will get ppc970s before any other vendor (although it's not sure if we're talking about high volume production here). this should confirm macbidouilles claim that the new systems are (or better: could at least - technically be) ready to ship.
mercury has also made some statements about the performance of their new systems (current machines use motorola g4 processors), they're talking about speed improvement up to 400%. but you should be careful here as those machines are all multiprocessor systems, and they're now using rapid-io as new bus system (using ppc970-to-rapidio bridges). as far as i know the new impact systems were originally designed to use mpc7500 processors, so mercury may have had the same problems as apple.
by the way: you should take a closer look to some mc circuits and compare them to the asic used in the xserve and the new powermacs. maybe apple and mc have closer relationships as some of you would think.
I doubt Mercury has a 970 product ready. From the ImpactRT 3100 product page:
"Harnessing PowerPC 7445 microprocessors to reach 600 GFLOPS or 2.4 TeraOPS, the ImpactRT 3100 multicomputer has the muscle to plow through the toughest computing tasks."
Companies like Mercury are rarely at the forefront of processor technology. It took them a year or so after the introduction of the G4 to get a product to market. I don't expect things to change with the 970.
While I don't know about hardware, Mercury does have a relationship with Apple in the software arena. They wrote vDSP. Mercury knows how to squeeze a proc for every last drop of performence when it comes to vector math.
Comments
Originally posted by mmicist
Ending sentences with prepositons, is, of course, grammatically wrong.
Should be : Unfortunately, that is the state in which our society is.
Churchill showed just how silly this can get with the marvellous phrase "This is something up with which I shall not put."
Hehe. I showed my wife (English minor, German major) my post as I sent it saying, "I wonder what people might say". I guess I should have taken a leaf out of the pages of The Inquirer's play book and turned my [irony alert] into, "Warning: Blatant grammatical error and attempt at humour!!".
MM
Originally posted by MartianMatt
Hehe. I showed my wife (English minor, German major) my post as I sent it saying, "I wonder what people might say". I guess I should have taken a leaf out of the pages of The Inquirer's play book and turned my [irony alert] into, "Warning: Blatant grammatical error and attempt at humour!!".
Subtlety, much like proper grammar and punctuation, is an art lost in online forums.
Originally posted by Kurt
I may have just missed his posts but I haven't seen Moki on these boards for a while. No comments on the MB rumors for dates and benchmarks. You don't suppose he has a test machine and is under a NDA do you?
It's fun to start rumors.
last time i mentioned that, he was very upset that i actually thought he had a ppc970 powermac to toy with.
he denied, offcourse
but who knows
Originally posted by Programmer
Subtlety, much like proper grammar and punctuation, is an art lost in online forums.
I sometimes worry that programmers code the way they write forum contributions. Punctuation, spelling and grammar do not seem to matter. Ah well!
Originally posted by Placebo
(repeat after me) Essence of language is comprehension! Essence of language is comprehension! Essence of language is comprehension! Essence of language is comprehension!
May I revise that to "Essence of language is unambiguous comprehension"? I've see far too many cases where both sides of a discussion come away "understanding" the outcome, but both "understanding" very different things.
Mr. MacPhisto
macrumors regular
Registered: Jan 2003
Location:
IBM Inside - some insider information
I've decided to post what I know about IBM's chip plans for the PPC and take some educated guesses as to how they relate to Apple.
Firstly, let me tell you why I know what I know. I have three good friends in the upper echelon of IBM. Two are in tech development positions of some kind, the other is in management. One is in Colorado, another in New York, and the last in North Carolina. I've gotten agreeing chip information from two of them. The third works in tech that doesn't involve chip manufacture. #3 has heard rumors, but hasn't been to official meeting or seen official documents.
Well, from what I understand, Apple came to IBM about two years ago requesting help because of Motorola's lack of interest in really building kick-butt processors for Apple. IBM agreed to help, although certain conditions were needed for this to happen. Most of this involved business related stuff and tech development costs, as well as some trade offs on technology developed by each company.
So, everyone knows about the 970 now, but there is also development on a new 750 based chip. There is already a 750 ready to go that is Altivec (I know, Moto terminology, just easier to know what it is that way) enhanced, can run with a FSB @ 400 or 450 MHZ with speeds up to 2GHZ. This is a modified 750fx, for all intents and purposes. These have been available for several months now, but Apple has not wanted them until now. It had to do with Apple's contractual obligations to Moto and the fact that they toast the G4 chip.
Apple is scheduled to have several palettes of these 750s (yes, up to 2 GHZ, lowest being 1.4) delivered to them in May. They also will have the 970s delivered to them @ speeds of 1.5 to 2.5. So, here's my educated guess:
The 970s now actually run more efficiently than when specs were announced, so that a high-end laptop could take a chip running at 1.8. I'm not sure if Apple will keep them all at 1.5 for power efficiency, but they may.
So, I think Apple introduces a completely revamped lineup in July. The iBooks and eMacs all use the 750 chip - iBooks running at 1.4, eMacs up to 2. The iMac may have a low-end model with this chip, and may be launched without the 970, giving the "Power" lineup a 3 to 6 month head start with the 970. I also don't think they'll call the 750 the G3. My bet is they may call it a Super G4 or a G4 Xtreme - something like that.
I see the PowerMacs getting an assortment of 970s, up to 2.5 with dual processor combos possible.
I think the 12" PB will get a 1.5 processor while the 15" PBs can get a 1.5 or 1.8 and the 17" comes with a 1.8
I think Apple may put 1MB of L3 on the 750 machines, 2MB on the PBs and PowerMacs. However, the PMs can take up to 4 - and it may be best for the laptops to only have 1.
I also know IBM is working on a 32bit new 750 based processor for Apple, so my guess is this will be used to delineate the iBook and eMac from the Power lineup. It may extend to the iMac too, but I'd guess a high-end iMac would have the 970.
And, now to the real kicker. Remember how I said that Apple had to make concessions to IBM? Well, one of them would allow IBM to manufacture and sell machines that ran Mac OS X - and would allow them sell it for servers, if a customer wanted it. This would not allow them to build laptops that ran the Mac OS, only desktop machines or servers.
Personally, I'm very skeptical, but hope some of it is true.
One mildly interesting idea in there, however, is that Apple wants to take legal action against Motorola but to do that they have to completely eliminate their dependency on Motorola products... and that they might hold back on the latest IBM chips until they can fullfill their contractual obligations and convert over completely. I'm skeptical, however. The timeframe on G3 w/ SIMD is aggressive and converting all entire lineups all at once seems too risky, especially when they just updated the eMacs. I don't buy it.
Originally posted by mmicist
Ending sentences with prepositons, is, of course, grammatically wrong.
Apologies if you're joking. If you're not joking, you're wrong.
English is not Latin. Honestly.
Originally posted by Programmer
Nothing really new here... the poster could simply have read this forum and put together a "rumour" that fullfills what people are speculating about.
One mildly interesting idea in there, however, is that Apple wants to take legal action against Motorola but to do that they have to completely eliminate their dependency on Motorola products... and that they might hold back on the latest IBM chips until they can fullfill their contractual obligations and convert over completely. I'm skeptical, however. The timeframe on G3 w/ SIMD is aggressive and converting all entire lineups all at once seems too risky, especially when they just updated the eMacs. I don't buy it.
Yes, the "Apple wants to take legal action" rumors seem a stretch to me. Let me rephrase that, maybe Apple would like to take legal action, but the rumors saying they intend to seem a stretch.
The part of this post that appears the most far fetched comes in the last paragraph(re: as other people posted in Mac Forums). It's possible that Apple would agree to allow IBM to manufacture desktops w/ Mac OS X, but I don't think it very likely. However, maybe Apple learned something in their foray into the clones and have included some more restrictions, nawww, I'm dreaming, ain't happening. Oh well, we'll soon find out.
Still hoping for the IBM 850, that was the hook this post got to me with.(preposition that
Originally posted by Nonsuch
Apologies if you're joking. If you're not joking, you're wrong.
English is not Latin. Honestly.
Ahhh, sanity ...
... in the end, good grammar is far more about good judgment, than it is about applying pre-set prescriptions: something that is completely lost in the way it's taught.
<back to regularly scheduled speculation>
Originally posted by tman
Macbidouille has stated that the new PM's will include Hypertransport. I haven't seen that addressed in this thread (although I'll admit that I haven"t read every post). Would someone much more knowledgeable than I comment on the ramifications of Hypertransport and the impact on throughput?
It depends on how it is used. The most likely use is to allow Apple to divide their chipset into two chips, one for I/O and one for the memory controller + FSB (+AGP?). That will help Apple rev the features of one without updating the other, and may allow them to share parts between high & low end lines.
Other uses of HT are harder to envision -- if they get a special GPU from ATI or nVidia (both of whom are part of the HT consortium) then it could replace AGP. This is much less likely, IMHO. The 970's FSB is HT-like but its not HT so this won't be used to communicate to the processor(s). Apple could also divide their chips further, but this tends to make the system more expensive so its not a benefit.
Really the use of HT is a technical detail that isn't particularly interesting to the end user.
Originally posted by Programmer
It depends on how it is used. The most likely use is to allow Apple to divide their chipset into two chips, one for I/O and one for the memory controller + FSB (+AGP?). That will help Apple rev the features of one without updating the other, and may allow them to share parts between high & low end lines.
Other uses of HT are harder to envision -- if they get a special GPU from ATI or nVidia (both of whom are part of the HT consortium) then it could replace AGP. This is much less likely, IMHO. The 970's FSB is HT-like but its not HT so this won't be used to communicate to the processor(s). Apple could also divide their chips further, but this tends to make the system more expensive so its not a benefit.
Really the use of HT is a technical detail that isn't particularly interesting to the end user.
The initial standard of HT was a daisy chain connection and the comunication between slaves in a daisy chain was not allowed. All comunication between slaves has to be done through master.
Later on the consortium has added the possibility of a switch like RAPIDIO.
If 970 has HT the communication between processors could be done with a switch.
I will try to provide a link.
regards
rooster
http://www.commsdesign.com/story/OEG20030127S0006
regards
rooster
mercury has also made some statements about the performance of their new systems (current machines use motorola g4 processors), they're talking about speed improvement up to 400%. but you should be careful here as those machines are all multiprocessor systems, and they're now using rapid-io as new bus system (using ppc970-to-rapidio bridges). as far as i know the new impact systems were originally designed to use mpc7500 processors, so mercury may have had the same problems as apple.
by the way: you should take a closer look to some mc circuits and compare them to the asic used in the xserve and the new powermacs. maybe apple and mc have closer relationships as some of you would think.
"Harnessing PowerPC 7445 microprocessors to reach 600 GFLOPS or 2.4 TeraOPS, the ImpactRT 3100 multicomputer has the muscle to plow through the toughest computing tasks."
Companies like Mercury are rarely at the forefront of processor technology. It took them a year or so after the introduction of the G4 to get a product to market. I don't expect things to change with the 970.
While I don't know about hardware, Mercury does have a relationship with Apple in the software arena. They wrote vDSP. Mercury knows how to squeeze a proc for every last drop of performence when it comes to vector math.